## [1] "Show relationships amongst variables visually"

## [1] "A-B scatter plot"
## [1] "A plot shows a level change from the 200th data point, can be controlled by a dummy variable"

## [1] "B plot looks fairly random/stochastic - avoiding tests of stationarity as the visual looks representative"

## [1] "C plot has a strong outlier at point 201"

## [1] "Autocorrelation for A - potentially two lags may be fitting, but none used as a useful model was found using the interaction effect (see below)"

#Removing outlier
content <- content[-c(201),]
#Creating Dummy Variable
content$D <- c(rep(0, 200), rep(1, NROW(content)-200))
#Modeling with interactions with Dummy given the break and phase shift in underlying series after 200th point
model <- lm(C ~ A * B * D, data=content)
summary(model)
##
## Call:
## lm(formula = C ~ A * B * D, data = content)
##
## Residuals:
## Min 1Q Median 3Q Max
## -125.899 -3.114 1.597 7.604 35.144
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 9.6114 1.6381 5.867 1.2e-08 ***
## A 1.3560 0.5124 2.646 0.00858 **
## B -5.3259 0.4310 -12.358 < 2e-16 ***
## D -2.3636 6.1447 -0.385 0.70077
## A:B 0.7792 0.1435 5.429 1.2e-07 ***
## A:D 0.9344 0.7797 1.198 0.23172
## B:D 22.2786 2.0070 11.100 < 2e-16 ***
## A:B:D -6.9617 0.2521 -27.619 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 16.69 on 291 degrees of freedom
## Multiple R-squared: 0.9465, Adjusted R-squared: 0.9452
## F-statistic: 735.5 on 7 and 291 DF, p-value: < 2.2e-16
#Residuals appear fairly stable
resi <- model$residuals
plot(resi)
